跳到正文
研踪PaperTrail

YOUR RESEARCH, IN SIGHT

让下一篇重要论文，不再错过。

覆盖计算机全学科，以 AI 为快捷入口。关注方向、作者和关键词，把发现变成持续的研究积累。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+0新增论文
+0其中 AI 相关
+2已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-08 07:11 · 全站最近入库：2026-11-08 06:02各信源进度 →

统计口径与数据新鲜度

统计 2026-11-07 07:31 — 2026-11-08 07:31（北京时间） · 点击数字查看对应简报；AI 相关包含在新增论文内。

最近完成检查：2026-11-08 07:11最近发现新增版本：2026-11-08 06:02库内论文最新更新时间：2026-11-06 02:59

每分钟刷新已保存的数据。空检查不代表采集异常；新收录可能包含历史论文。AI 导读按新增版本计数。

17,100 已收录论文14,661 AI 相关4,057 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

时间范围跟随排序：版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常，查看状态 →

找到 17,100 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-08 07:09（北京时间）
版本更新 · v2arXiv:2610.37466

Demystifying the Impact of Physical Non-Idealities on Kernel Performance

Supun Kuruppu, Pranav Dangi, Wei Li, Zhenyu Bai, Tulika Mitra

原文摘要

Physical non-idealities in integrated circuit (IC) design, fabrication, and operation are concealed by conventional hardware abstractions, causing unexpected performance variations in kernels. For large-scale workloads, such as AI training and inference, performance unpredictability can lead to substantial losses. We investigate this problem in NVIDIA H100 GPUs with their distributed shared memory (DSMEM) architectural feature that reduces data movement from the L2 cache to Streaming Multiprocessors (SMs) with multicast loads, and uncover that floorsweeping restricts the number of SMs usable for clustered execution, increasing the number of execution waves of kernels. To alleviate this performance penalty, we introduce the wave-aware clustering strategy that accounts for these resource constraints and improves the throughput of the Hopper version of FlashAttention 4 by up to 5.30% over the unclustered baseline. Furthermore, we discover that the power constraint imposed by the thermal design power (TDP) limit causes the DSMEM-based memory bandwidth optimization to improve the throughput even for compute-bound kernels by creating a power headroom for higher SM frequencies through reduced data movement. Extending this analysis beyond GPUs, we find that power constraints also limit throughput scaling on Tenstorrent's Wormhole accelerator, highlighting the broader importance of accounting for physical non-idealities in kernel performance engineering.

cs.AR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04296

SEAM: Situated Executable Associative Memory for a Co-located Community

Keru Wang, Michael Xu, Ken Perlin, Nebojsa Jojic, Andrzej Banburski-Fahey

原文摘要

In groups, knowledge is distributed among individuals, and how it is brought together depends on social ties. Taking inspiration from how a single mind recalls and reacts to memories and current stimulus, we build a shared memory that does this for a community, prototyped as a social app, SEAM. Members post arbitrary moments; SEAM works out how fragments relate, distills knowledge about the group, and replies in a situated and associative way. Some of these memories act as instructions that can alter how the system behaves, making it programmable by its members. We deployed SEAM to 46 people in a co-located community for two weeks of open-ended use, and ran a follow-up workshop with 14. Members shared their days, interrogated the memory, and instructed it to track routines and run games; replies introduced them to people and subjects they had never encountered. SEAM points towards communities where knowledge finds people.

cs.HC · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.19691

Efficient construction of subspace design codes over constant alphabet

Shu Liu

原文摘要

Subspace-design codes form a central unifying framework in coding theory, connecting probabilistic existence results and explicit algebraically structured constructions. The two core families are folded Reed-Solomon codes and univariate multiplicity codes, both possessing the subspace-design property. These codes reach list-decoding capacity with polynomial-size lists, but their alphabet size grows with code length and hurts practical efficiency. Goyal, Guruswami, and Hsieh achieved a major breakthrough: they constructed the first explicit subspace-design codes over constant-size alphabets via the Alon-Edmonds-Luby expander framework, settling an open problem and showing subspace-design properties can hold without large alphabets. Despite recent progress, constructing subspace-design codes with low-complexity efficient encoding remains a critical open challenge. This work gives an explicit construction of r-dimensional subspace-design codes over a constant alphabet whose size grows exponentially with r. By comparison, Goyal et al. (FOCS 2026) built r-dimensional subspace-design codes with alphabet size doubly exponential in r squared. Our construction is computationally efficient: encoding scales quadratically with block length and linearly with dimension r in field operations. We use modified algebraic geometry codes, replacing rational evaluation points with higher-degree places to obtain a valid additive code. To confirm parameter bounds, we use Moore matrices to analyze determinants with many zero entries. We also generalize univariate multiplicity codes, originally built from Reed-Solomon codes, to algebraic geometry codes, and prove these AG-multiplicity codes achieve optimal parameters.

cs.IT

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04295

Reciprocity Can Erase Direction:tructural Boundaries and Controlled Utility of Magnetic Graph Networks

Dawei Han, Zicheng Wang, Yaoxing Ren, Yi Lv, Rongqian Wu, Fumin Wang

原文摘要

Magnetic graph networks encode direction through phase, but a learned graph can erase the intended directional mechanism before prediction. We characterize this failure for support-based magnetic Laplacians. Exact identities relate imaginary energy and distance from zero phase to amplitude-weighted reciprocity. Top-$k$ selection forces complete collapse when the observed-node count satisfies $m\leq k+1$, and dense support imposes a lower bound on reciprocal-arc fraction. We then construct paired propagation examples with identical node features and symmetric amplitudes but opposite labels, giving a provable chance-level limit for direction-erasing representations. In 72 fixed neural fits, support phase achieved test AUROC 1.0000 on a unilateral ring but 0.5000 on reciprocal ring and complete supports; continuous score phase and a real directed comparator achieved 1.0000 throughout. All models solved the direction-free control. These are oracle-direction mechanism tests, not evidence of learned clinical direction. In the preserved public DeepDRiD development study, continuous phase was active but its robustness difference from zero phase was only $+0.0036$ (descriptive 95\% interval $-0.0014$ to $+0.0095$) on validation patients used for checkpoint selection. The contribution is a structural and experimental framework for distinguishing erased direction from unestablished task utility. It provides a precise circumstance in which magnetic phase preserves useful information, without asserting universal or clinical superiority.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04294

State Evolution and Universality of Iterated Linear Programs for the Binary Perceptron

Jianan Zhang, Shuangping Li

原文摘要

The binary perceptron is a random constraint satisfaction problem that asks for a sign vector satisfying a system of random linear inequalities. Its solution space has a strikingly extreme geometry: at every positive constraint density, all but a vanishing fraction of the solutions are isolated. Yet polynomial-time algorithms find solutions at small enough constraint density, which raises the question of how efficient algorithms navigate such a landscape. Linear programming is a natural algorithm to study here, but its analysis has so far been limited to a single round. Li, Schramm, and Zhou (STOC 2025) analyzed one linear program followed by iterative rounding. In this work, we overcome this limitation and give a precise analysis of iterated linear programming for the binary perceptron. (1) For Gaussian matrices and any fixed number of rounds, we characterize the iterated linear programs exactly through a deterministic state-evolution recursion. (2) Adding a final rounding step, we obtain a randomized polynomial-time algorithm for the binary perceptron. Numerical evaluations of the recursion provide evidence that it succeeds at constraint density $α=0.60$ with zero margin, well beyond the range $α\le0.1$ guaranteed by the single-LP analysis and $α\approx0.327$ (Jo and Lee, 2026), the best guarantee so far. (3) We establish universality, showing that the same characterization holds for matrices with i.i.d. entries drawn from any centered, unit-variance subgaussian law. The key idea of our proof is that, although conditioning on earlier rounds makes the matrix non-Gaussian, its conditional law can be computed explicitly and remains close to a Gaussian law, so the convex Gaussian min-max theorem applies in every round. For universality, we compare Gaussian and subgaussian matrices through the Lindeberg principle, using a regularized version of each linear program.

cs.DS · cs.CC · math.PR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04292

How Risky Is Your Data-Driven Policy?

Omar Besbes, Alon Rieger

原文摘要

Data-driven algorithms are ubiquitous across industries, and there is an extensive literature on analyzing their \emph{average} performance. The risk profile of such algorithms, while critical, is much less understood. This paper quantifies the risk of data-driven policies and isolates its sources. We evaluate a policy through the conditional value-at-risk (CVaR) of its loss over the training sample, any policy randomization, and the future outcome, at a fixed unknown distribution. Regret against an oracle that knows the distribution decomposes into expected post-sample regret (PS), the average outcome-only CVaR gap of the selected action, and a sample-risk premium (SR), the additional tail risk omitted by averaging those evaluations. The analogous premium vanishes under expected loss. We characterize when the premium vanishes and how large it is relative to post-sample regret. For losses with a continuous density near the tail cutoff, we identify conditions under which both components decay at the same rate, yet the limiting ratio of the premium to post-sample regret can reach the thousands. When a probability mass at the cutoff moves with the action, the premium can decay more slowly and dominate post-sample regret. We establish conditions under which sample average approximation learns the oracle action at rate $n^{-1/2}$ in both settings, but its combined-risk regret is of order $n^{-1}$ in the first and $n^{-1/2}$ in the second. In finite-sample numerical examples of SAA in both settings, the premium accounts for approximately $74\%$--$97\%$ of regret.

stat.ML · cs.LG · math.OC

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2610.28569

ECHO: Efficient Cascade with Homogeneous Operators for Distillation-Guided 100K-Sequence Modeling

Wenhao Li, Qinglin Jia, Chenghou Jin, Zihong Xian, Jiezhou Ji, Yanzhi Liu, Zezheng Wang, Bo Chen 等 12 位作者

原文摘要

Modeling lifetime user behavior sequences is essential for capturing long-range preferences in industrial recommendation, yet sequences containing up to 100K interactions make target-aware ranking prohibitively expensive. Search-based methods that retrieve from raw sequences incur candidate-dependent costs that grow with sequence length, whereas latent compression methods reduce computation at the expense of item-level addressability. We propose ECHO (Efficient Cascade with Homogeneous Operators), a framework that balances efficiency and fine-grained retrieval by retaining a subset of original event records per user through a Distilled Selector (DS) and retrieving a smaller set of target-relevant records for each candidate through a Target-aware Fine-grained Retriever (TFR). DS is supervised by a candidate-aware full-history teacher via per-target distillation, enabling target-aware user-level selection without requiring full-history computation for every candidate. The retriever and target-attention scorer share the relevance parameters learned by the teacher, ensuring consistent relevance estimation throughout the cascade. At a sequence length of 100K, ECHO reduces online sequence-branch FLOPs by approximately 160 times compared with a representative end-to-end model. Three deployment optimizations, including common user inputs, compact record encoding, and key-side projection absorption, further reduce memory, storage, and data-movement overhead. Experiments on KuaiRand-1K and an industrial-scale short-video platform demonstrate relative GAUC improvements ranging from 0.113% to 1.520% over state-of-the-art baselines across multiple tasks. Furthermore, online A/B tests in two production scenarios, Kuaishou and Kuaishou Lite, show that ECHO improves app usage time by 0.240% and 0.373%, respectively.

cs.IR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04291

The Native Menace: Host-to-Browser Cross-Boundary Attacks via Native Messaging

Jyotirmay Chauhan, Rownak Islam, Jason Polakis

原文摘要

Modern browsers are complex software platforms that users can customize through browser extensions to further expand their functionality. Extensions derive their functionality from privileged browser APIs that expose powerful capabilities. Prior work has extensively studied how malicious webpages or extensions can target those capabilities. However, prior research has overlooked the NativeMessaging API, which allows extensions to create a communication channel that bridges the gap between web content rendered in the browser and native apps running on the user's machine. This API enables workflows that cross the browser's boundary to perform critical operations (e.g., digital signing) or exchange sensitive data (e.g., credentials). In this paper, we conduct the first, to our knowledge, empirical security analysis of NativeMessaging, and present a novel attack that allows a non-privileged malicious native app to misuse browsers' native messaging capabilities to stealthily intercept sensitive communications or launch confused deputy attacks against other native apps. To demonstrate the significant risk introduced by native messaging, we present a series of concrete attacks against popular and critical apps (e.g., 1Password, the European Union's eID system, Google Drive), whose browser extensions have more than 532 million cumulative downloads. Our attacks have severe ramifications, including full password vault leakage, and arbitrary document signing in governmental workflows. Fundamentally, these attacks are possible due to native messaging being built on top of an untrusted communication channel without providing a root of trust or enforcing some form of client attestation. Accordingly, we propose a straightforward mitigation technique of message encryption with out-of-band secret sharing, to prevent our attacks.

cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04289

When Debate Amplifies Risk: Social-Bias Vulnerabilities in Multi-Agent Safety Deliberation

Qinghua Mao, Xi Lin, Yuxin Qi, Zihao Xu, Jinze Gu, Yuliang Chen

原文摘要

Retrieval-augmented Web applications increasingly employ multi-agent debate (MAD) to aggregate conflicting, noisy, and heterogeneous information, yet deliberation is fundamentally a social process susceptible to interpersonal dynamics. When retrieved Web sources contain latent hazards alongside safe factual content, social biases such as conformity pressure and authority deference can override evidence-grounded safety boundaries. In this work, we investigate whether the social structure of deliberation can paradoxically amplify latent Web safety risks. We introduce SafeConflict-Bias, a controlled counterfactual diagnostic protocol spanning six core Web risk categories, which holds queries and retrieved evidence strictly invariant while isolating conversational majority sway, authority cues, and confidence asymmetry. Empirical evaluation across diverse proprietary and open-source backbones reveals that while isolated models reliably reject unsafe guidance, standard MAD catastrophically capitulates to social pressure---surging unsafe adoption from 0.0\% up to 100.0\% and systematically silencing safe minority evidence. To resolve this systemic fragility without retraining or debate regeneration, we propose RiskTrace-Debias, an inference-time auditing framework uniting evidence-first precommitment, social-metadata masking, and consistency-gated arbitration. RiskTrace-Debias suppresses worst-case unsafe adoption to $\le 2.0\%$ across all architectures without inducing over-refusal, while generating verifiable provenance traces to audit interpersonal manipulation in Web deliberation.

cs.CR · cs.AI · cs.CY

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04286

GroundSeek-CXR: Retrospective Evaluation-Guided Active Observation Framework for Multi-Label Chest Radiography Diagnosis

Yihang Fu, Xinyu Liu, Qing Xu, Xiaofeng Yang, Zhen Chen

原文摘要

Chest radiography (CXR) diagnosis is critical for thoracic disease assessment, yet challenged by the need to identify subtle lesions and co-occurring abnormalities across regions. While recent vision-language models can actively observe local regions, existing preference-based methods supervise each observation by its immediate effect on the current diagnosis, failing to capture how each observation and stopping decision contributes to the final diagnosis. To address this issue, we propose GroundSeek-CXR, an active CXR diagnosis framework that learns observation and stopping decisions by retrospectively evaluating them with the final diagnosis. Specifically, we first devise a Retrospective Diagnostic Evaluation (RDE) that scores observation and stop candidates from offline Monte Carlo tree search through independent continuations. Moreover, we design a Retrospective Gap-weighted Preference Optimization (RGPO) that learns from retrospective-score preferences weighted by their score gaps, prioritizing actions that improve diagnostic accuracy. Extensive experiments demonstrate that our GroundSeek-CXR framework outperforms state-of-the-art methods, and improves abnormal-image sample-F1 by 4.48% over immediate-score preference learning.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04285

Optimal Communication for Distributed Sparse Histogram Aggregation

Fangwei Ye, Yijun Fan

原文摘要

Histogram estimation is a fundamental primitive in distributed data analysis, with applications in counting queries and federated learning. In many settings, the universe of categories is enormous, yet each user's data involves only a small subset, yielding sparse local histograms. This raises a natural question: To what extent can local sparsity reduce the communication needed to estimate the global histogram? We answer this question by characterizing the communication complexity of distributed sparse histogram estimation under prescribed $\ell_1$ distortion and failure probability constraints. Our upper bound is achieved by a randomized covering scheme that optimizes how accurately each user represents its local histogram to reduce communication. By exploiting the distortion budget directly rather than imposing unbiasedness, the scheme achieves lower communication costs than standard unbiased sampling and hashing constructions. We establish the scheme's first-order optimality through a lower bound based on local support identification. The key step is to relate communication cost directly to the number of active coordinates revealed by each user's message, rather than to bound information about the aggregate via distance-based Fano's inequalities. This characterization further quantifies how the communication advantage of heterogeneous over homogeneous encoders varies across parameter regimes.

cs.IT

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04284

OmniScribble: Mixed-Initiative Authoring of Animated Sketches through Inspectable and Repairable AI Decisions

Hien Trinh, Tan-Hiep To, Trung-Nghia Le

原文摘要

AI-assisted creative tools reduce the effort of authoring content, but delegating creative decisions to AI can make it difficult for authors to understand, correct, or redirect work. We present OmniScribble, a system for animating multi-object sketches that treats object decomposition and motion planning as editable artifacts. Rather than generating animation end-to-end, OmniScribble presents AI proposals at verification checkpoints, where authors can accept, repair, or regenerate them. Both manual and AI-assisted authoring operate on the intermediate representation and rendering pipeline, allowing us to isolate AI delegation. In a within-subject study with eight participants, AI-assisted authoring yielded a median within-participant reduction in completion time of 71.4\%, while perceived control and motion-intent alignment remained high. Interaction traces and interviews show that participants adapted to AI errors by repairing local issues, regenerating mismatches, and accepting proposals. These findings suggest that inspectable, repairable delegation enables generative automation while maintaining control over the outcome.

cs.HC

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2610.30170

Accelerating Wildfire Hazard Mapping Across Canada with an AI Surrogate Modelling Framework

Hager Radi Abdelwahed, Toluwaleke Olutayo, Chirag Goel, Jana Osea, Aldo Zaimi, Tom Swystun, Sandy Erni, Francisco Emiliano Lopez Saavedra 等 10 位作者

原文摘要

Canada's increasing wildfire activity requires frequent national hazard mapping to guide proactive mitigation, including fuel management, land-use planning, infrastructure protection, and evacuation routing. However, process-based wildfire simulation frameworks (e.g., BurnP3 and BurnP3+) rely on computationally intensive Monte Carlo simulations to model burn probability, fire intensity, and rate of spread. These workflows can take days to weeks to process a single regional modelling unit. To overcome these computational bottlenecks, we present three contributions developed through collaboration between AI researchers and wildfire modelling experts: (a) a standardized framework for training, fine-tuning, and evaluating AI surrogate models against process-based wildfire simulators; (b) a multi-task deep learning surrogate trained on national-scale data to emulate burn probability, mean fire intensity, and mean rate of spread at a fraction of the computational cost (seconds); and (c) domain-specific metrics and counterfactual evaluation protocols to probe out-of-distribution generalization. Integrating geospatial data (fuel, terrain, ignition likelihood) and regional weather, our model approximates BurnP3+ simulation outputs, achieving strong performance across agreement, rank-correlation, error, and spatial-overlap metrics. It preserves downstream hazard classifications and produces physically plausible responses to expert-informed counterfactual interventions. Furthermore, we demonstrate zero-shot transfer to a separate regional BurnP3+ study and show that lightweight fine-tuning can substantially improve regional calibration and predictive agreement. On a representative region, the surrogate achieves 806-fold end-to-end CPU acceleration and uses 13% of BurnP3+'s host memory, facilitating rapid scenario analysis.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04283

Metric TSP in Sublinear Time: Breaking Two Approximation

Amir Azarmehr, Soheil Behnezhad

原文摘要

The seminal result of Czumaj and Sohler [STOC'04] showed that a 2-approximation of the value of metric TSP can be obtained with sublinear queries. Whether sublinear algorithms can break the 2-approximation barrier has remained a major open problem. See problem 71 on sublinear.info. This paper resolves this open problem. We show there are fixed $\varepsilon, δ> 0$ such that a $(2-δ)$-approximation of metric TSP value can be obtained with $O(n^{2-\varepsilon})$ queries. Here $n$ is the number of vertices, and the input size is $Θ(n^2)$.

cs.DS

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.30350

DynaFLIP: Rethinking Robotics Perception via Tri-Modal-Dynamics Guided Representation

Jusuk Lee, Seungjae Lee, Jonghun Shin, Hoseong Jung, Sungha Kim, Daesol Cho, H. Jin Kim, Jia-Bin Huang 等 9 位作者

原文摘要

Robot manipulation critically depends on perception that preserves the action-relevant aspects of a scene. Yet most robot learning pipelines are built upon visual encoders pre-trained for static recognition or vision-language alignment, leaving motion understanding to downstream policies. We introduce DynaFLIP, a dynamics-aware multimodal pre-training framework that pushes motion understanding upstream into perception. We construct image-language-3D flow triplets from heterogeneous human and robot videos, and use these triplets as training-time supervision to shape an image-only encoder. Our key idea is to encourage the three modalities to span a small simplex volume in the shared hyperspherical space -- a smaller simplex volume indicating stronger alignment. To avoid the geometric ambiguity and trivial collapse of naive volume minimization, we combine simplex-volume minimization with a cosine regularizer and a contrastive objective. Our analyses show that DynaFLIP focuses on control-relevant regions critical for manipulation. The resulting dynamics-aware representations serve as reusable visual backbones and consistently outperform baselines across diverse downstream policies, including VLAs. We validate this across diverse simulation and real-world setups, with gains reaching +22.5% under out-of-distribution scenarios. Our results suggest that robot generalization improves when visual representations are trained to encode not just what is present, but how the world changes under action.

cs.RO · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2604.10784

TorchUMM: A Unified Multimodal Model Codebase for Evaluation, Analysis, and Post-training

Yinyi Luo, Wenwen Wang, Hayes Bai, Hongyu Zhu, Hao Chen, Pan He, Marios Savvides, Sharon Li 等 9 位作者

原文摘要

Recent advances in unified multimodal models (UMMs) have led to a proliferation of architectures capable of understanding, generating, and editing across visual and textual modalities. However, developing a unified framework for UMMs remains challenging due to the diversity of model architectures and the heterogeneity of training paradigms and implementation details. In this paper, we present TorchUMM, the first unified codebase for comprehensive evaluation, analysis, and post-training across diverse UMM backbones, tasks, and datasets. TorchUMM supports a broad spectrum of models covering a wide range of scales and design paradigms. Our benchmark encompasses three core task dimensions: multimodal understanding, generation, and editing, and integrates both established and novel datasets to evaluate perception, reasoning, compositionality, and instruction-following abilities. By providing a unified interface and standardized evaluation protocols, TorchUMM enables fair and reproducible comparisons across heterogeneous models and fosters deeper insights into their strengths and limitations, facilitating the development of more capable unified multimodal systems. Code is available at: https://github.com/AIFrontierLab/TorchUMM.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2603.18373

To See or To Please: Uncovering Visual Sycophancy and Split Beliefs in VLMs

Rui Hong, Shuxue Quan

原文摘要

When Vision-Language Models (VLMs) answer correctly, do they rely on the image? We introduce a Tri-Layer Diagnostic Framework that asks three questions of every (image, question) pair when the image is blanked: does the model register that the image is missing (Latent Anomaly Detection, LAD), does its output distribution depend on the image (Visual Necessity Score, VNS), and does it nonetheless commit to a content answer? The cascade assigns each pair of any existing benchmark to one of four modes: Perceptual Blindness, Language Shortcut, Visual Sycophancy (the blank image registers and matters, yet the model commits to a fabricated answer), or Robust Refusal. Across nine open-weight VLMs and 9,000 model-sample pairs, 28.1% are Visual Sycophancy, 18.6% Language Shortcut, 5.6% Perceptual Blindness, and 47.8% Robust Refusal. The modes are strongly model-specific: the Qwen3-VL models refuse on blank input in 93-98% of pairs, whereas Pixtral-12B commits to an answer in 72%. Within-generation scaling increases Visual Sycophancy in both Qwen2.5-VL (7B->72B: 25.5->40.1%) and Llama-3.2-Vision (11B->90B: 34.3->48.0%), while the newer Qwen3-VL generation nearly eliminates it. A blinded human study confirms that LAD and VNS track what annotators observe, and that 80% of Visual-Sycophancy responses carry no textual sign that the image was blank.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04282

TTSonic: A Unified Framework for Composable and Controllable Speech Generation

Zhedong Zhang, Kaiwen Zhou, Junyan Wu, Shengqiang Li, Xuran Chi, Xingui Zeng, Rui Yan, Xu Chen

原文摘要

Modern text-to-speech (TTS) systems can synthesize natural and expressive speech, yet a practical speech foundation model must go beyond naturalness to offer stable quality, precise controllability, and a broad set of composable capabilities within a unified framework. We present TTSonic, a unified speech generation model that supports zero-shot TTS, voice cloning, prosody cloning, instruction following, voice conversion, and cloning from noisy reference speech. Rather than relying on separate task-specific models, TTSonic accommodates diverse conditioning signals, including text, natural language instructions, reference speech, and text-described timbre, within the same model framework. Specifically, we develop a semantic--acoustic disentangled speech tokenizer through ASR-supervised semantic learning and reconstruction-guided generative refinement. Serving as the representational foundation of our unified framework, the tokenizer enables the Text-to-Semantic autoregressive model to focus on semantic and prosodic planning, while the acoustic flow-matching model specializes in high-quality speech rendering under diverse acoustic and timbre conditions. Building on this framework, we incorporate CoT-guided instruction-following planning and Text2Timbre voice design to support controllable speech generation. Experiments across zero-shot synthesis, multilingual and cross-lingual TTS, instruction following, voice design, voice conversion, and noisy-reference cloning demonstrate the competitive performance and broad applicability of TTSonic.

cs.SD

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2603.28458

HISA: Efficient Hierarchical Indexing for Fine-Grained Sparse Attention

Yufei Xu, Fanxu Meng, Fan Jiang, Yuxuan Wang, Ruijie Zhou, Zhaohui Wang, Jiexi Wu, Zhixin Pan 等 14 位作者

原文摘要

Token-level sparse attention mechanisms, exemplified by DeepSeek Sparse Attention (DSA), achieve fine-grained key selection by scoring every historical key for each query through a lightweight indexer, then computing attention only on the selected subset. While the downstream sparse attention itself scales favorably, the indexer must still scan the entire prefix for every query, introducing an per-layer bottleneck that grows prohibitively with context length. We propose HISA (Hierarchical Indexed Sparse Attention), a plug-and-play replacement for the indexer that rewrites the search path from a flat token scan into a two-stage hierarchical procedure: (1) a block-level coarse filtering stage that scores pooled block representations to discard irrelevant regions, followed by (2) a token-level refinement stage that applies the original indexer exclusively within the retained candidate blocks. HISA preserves the identical token-level top-sparse pattern consumed by the downstream Sparse MLA operator and requires no additional training. On kernel-level benchmarks, HISA achieves up to speedup at 64K context. On Needle-in-a-Haystack and LongBench, we directly replace the indexer in DeepSeek-V3.2 and GLM-5 with our HISA indexer, without any finetuning. HISA closely matches the original DSA in quality, while substantially outperforming block-sparse baselines.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.01837

Computons for Control-Driven Boolean Circuits

Damian Arellanes

原文摘要

Boolean circuits abstract away from hardware aspects to capture the logical structure and computational behaviour of digital components. Despite their long history of study, their compositional foundations have received comparatively little attention. Building on the theoretical foundations of computons introduced in previous work, we develop a concrete and operationally defined model of high-level computation for control-driven Boolean circuits. We describe colimit-based operators for composing such circuits into sequential, parallel, branching and iterative structures, supported by concrete operational semantics grounded in abstract computing devices. The framework refines the semantics of composition, establishes the algebraic properties of its operators and introduces atomic circuits to support more expressive computation units. We demonstrate the applicability of the proposed framework through examples from propositional logic, biomedicine and electronics, and establish an operational equivalence with respect to classical Boolean circuits through a systematic transformation that accommodates gates of arbitrary arity. By this equivalence, we show that control-driven Boolean circuits are at least as powerful as their classical counterpart. That is, they are capable of non-uniformly computing any Boolean function over inputs of arbitrary length. Implementations in Idris and Java validate the practical realisability of the framework.

cs.LO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04278

Unmitigated? How are the world's largest companies responding to AI risks?

Alexander K. Saeri, Jess Graham, Michael Noetel, Peter Slattery, Kun Zhao, Eunice Zhang, Erin Howard, Neil Thompson

原文摘要

AI is introducing new risks to large companies and amplifying existing risks, but whether companies are addressing the AI risks they are exposed to is largely untested. We investigated what outsiders can observe: whether companies publicly acknowledge AI risks. We coded public documents from 200 of the largest companies worldwide, collected in early 2026, against the 24 AI risks in the MIT AI Risk Repository, and compared acknowledgements with expert judgements of risk severity, sector vulnerability, and AI role responsibility from a companion Delphi study of 272 AI risk experts. Of the 100 largest companies, 32% did not publicly acknowledge any AI risks. Of the largest 1000 companies, we estimate that 62% [95% CI: 54%, 70%] did not publicly acknowledge any AI risks. Where companies did acknowledge risks, five accounted for over half of all acknowledgements: discrimination, weapons & cyberattacks, fraud & scams, loss of privacy, and devaluation of human creativity. Several risks that experts judged most severe were almost absent: almost no company acknowledged power centralization (concentration of power among those who control advanced AI), competitive dynamics (racing to deploy AI at the expense of safety), or dangerous capabilities. Companies in sectors experts judged more vulnerable to a risk were more likely to acknowledge that risk. But even companies in sectors judged 'extremely vulnerable' to a risk usually did not acknowledge it. Companies with roles experts judged most responsible for addressing a risk were no more likely to acknowledge it than other companies. General-purpose AI developers acknowledged AI misalignment (15%) far less often than discrimination (85%). These findings provide an empirical baseline for the design of AI risk disclosure requirements.

cs.CY

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04277

Harness-S: Learning from Sibling Harnesses for Generalizable Agent Training

Kangning Zhang, Wenxiang Jiao, Bowen Fang, Shuai Shao, Jiachen Zhu, Rong Shan, Shijian Wang, Hao Wang 等 11 位作者

原文摘要

Harness evolution from a common base exhibits Locality, Complementarity, and Nonstationarity: variants repair different failures, and their utility changes as the policy learns. Rather than seeking a universally superior harness, we introduce Harness-S to selectively rescue failures of a base execution system through sibling harnesses and transfer the resulting capabilities into a single policy that remains effective under the original harness. Directional Harness Evolution develops specialized sibling harnesses along distinct repair directions, forming a reusable Sibling Harness Pool that remains fixed during policy training. Complementary Harness Rescue Training invokes this pool only when all initial attempts under the base harness fail. The same policy generates fresh rollouts under selected auxiliary harnesses, and these trajectories replace the failed group for reinforcement learning. An Adaptive Harness Router selects the Top-k variants using Harness Behavior Profiles and voting across multiple LLMs. Policy-aware profile re-anchoring refreshes this evidence as the policy learns. Across AppWorld, SWE-Gym, and DBBench, the trained policy substantially outperforms policy-only training when returned to the original base harness. Behavioral analyses and held-out cases support the transfer of harness-elicited behaviors into the policy itself. Reusing the router for Rank-1 inference provides additional gains at the same executor rollout count.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04276

Testing Implies Distance Estimation with Polynomial Loss

Gaia Carenini, Cameron Seth

原文摘要

A tester decides whether a graph has a property or is far from having it. A distance estimator asks for more: it must approximately determine how many edges need to be changed. A property is said to be testable if there is a tester with a number of queries bounded independently of the graph's order for each conceivable distance. Similarly, if a property admits an estimator with a number of queries bounded independently of the graph's order, it is said to be estimable. Fischer and Newman proved that, in the dense graph model, every testable property is also estimable. We show that the passage from testing to estimation requires only a polynomial increase in query complexity. This resolves a long-standing open problem in property testing, previously tackled by Fiat and Ron, by Hoppen, Kohayakawa, Lang, Lefmann, and Stagni, and by Gishboliner, Kushnir, and Shapira.

cs.DS · math.CO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04275

Online Metric Matching with Unknown i.i.d. Arrivals

Tsubasa Harada

原文摘要

We study online metric matching with $n$ servers and $n$ requests drawn independently from an unknown distribution. Each request must be matched immediately and irrevocably to an unused server, and the algorithm receives no samples before the process starts. We give a randomized algorithm that is $O((\log\log\log n)^2)$-competitive on arbitrary metric spaces and $O(1)$-competitive on tree metrics. These guarantees match the corresponding bounds for known i.i.d. request distributions in [GGPW19] up to constant factors. More generally, our competitive ratio is $O(α(n))$ whenever the Fair-Bias algorithm introduced in that work has competitive ratio $α(n)$ for known i.i.d. arrivals on the same class of metrics. Consequently, proving that Fair-Bias is constant-competitive on general metrics would also yield a constant-competitive algorithm for unknown i.i.d. arrivals.

cs.DS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04274

Reasoning about Shared-Qubit Concurrency in Quantum Computing

Mingsheng Ying

原文摘要

In this paper, we develop a semantic model for the parallel execution of quantum circuits, providing a formal foundation for studying shared-qubit concurrency in quantum computing. As in classical shared-memory concurrency, parallel quantum circuits may interfere with one another through operations on shared qubits. However, quantum entanglement and the noncommutativity of quantum operations introduce additional challenges for reasoning about their correctness. To address these challenges, we generalise the classical Owicki-Gries method to quantum computing at the level of quantum circuits. Our approach enables compositional reasoning about parallel quantum circuits by separating the verification of individual components from the analysis of interference between them. We formulate proof rules for establishing the correctness of parallel quantum circuits and prove their soundness with respect to the proposed semantic model.

quant-ph · cs.DC · cs.LO · cs.PL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04273

Deep Learning Compilers: Architecture, Optimization, and Empirical Characterization

Jinghao Wen, Ruixuan Wang, Xun Jiao

原文摘要

As deep learning (DL) models become increasingly complex and hardware architectures grow more heterogeneous, DL compilers play a critical role in optimizing computational efficiency and enabling scalable deployment. This paper presents a systematic analysis of DL compilers, covering their architectural evolution, compilation workflows, optimization techniques, and emerging research directions. We organize DL compiler architectures into frontend, middle-end, and backend stages, emphasizing the role of the middle-end in connecting high-level model representations with hardware-specific optimizations. We further examine representative DL compilers and recent optimization techniques across different compilation stages. Beyond architectural and technical analysis, we conduct an empirical evaluation of compiler behaviors to investigate the effectiveness of graph optimization, compilation overhead, and dynamic-shape handling. Experiments using PyTorch/TorchInductor on an NVIDIA RTX A6000 demonstrate that compiler benefits depend strongly on workload characteristics, with speedups of up to 8.21\times for fusion-friendly computation graphs, while other workloads exhibit limited improvements. Our results further reveal that compilation overhead can require thousands of executions to amortize, and dynamic-shape compilation reduces cumulative shape-transition costs by approximately 70\% at the expense of modest steady-state performance degradation. By combining systematic architectural analysis with quantitative experimental evidence, this work provides insights into the design, optimization, and practical deployment of modern DL compilers.

cs.LG · cs.AR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04271

From Guidelines to Practice: Defining Requirements for AI Red-Teaming Tools

Gia Phong Tran, Thai Duong Huynh, Randolph Loh, Balachandar Gowrisankar, Zihao Wang, Moe Thu Kyaw, Jun Sun, Adams Wai-Kin Kong 等 9 位作者

原文摘要

Artificial Intelligence (AI) has seen significant developments and increased usage in recent years. AI benefits various tasks, automates decision-making, and enhances efficiency. However, AI also introduces additional attack surfaces for threat actors to perform attacks like prompt injection and jailbreaking, making security a great concern. AI security testing aims to imitate adversarial attacks or malicious inputs via manual testing or with the assistance of tools. Thus, it is critical that organizations are aware of the capabilities and limitations of the tools before using them. Existing studies provide insights on the overall architecture of AI security testing tools and raise concerns on their reliability, but their analysis is limited to attack performance. Selection of a tool for AI risk assessments should also depend on their alignment with the recommendations of industry standards, but to our knowledge, previous studies have not utilized such guidelines. To bridge this gap, we analyze several industry standards and guidelines to identify the requirements of AI testing tools. We discuss and highlight desirable features across various AI security testing tools for a variety of different AI systems. We evaluated six tools across six Large Language Models, reported on our insights, and identified gaps that should be considered by future tools.

cs.CY

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2604.20857

DiagramBank: Draw from the Great, Draw Your Best

Ling Yue, Jiaying Wang, Zhexi Lu, Mingzhi Zhu, Tingwen Zhang, Zhen Xu, Shaowu Pan

原文摘要

Scientific diagrams help researchers communicate complex methods and relationships, but creating them requires substantial effort in content organization, layout design, and visual refinement. Automated diagram generation can reduce this workload. Reference examples can guide this process by showing how scientific ideas are organized and presented visually. Diagrams from accepted papers provide a rich source of such examples, but their utility as a large-scale reference resource remains underexplored. We introduce DiagramBank, a diagram-text reference resource containing 57,100 scientific diagrams from 26,751 accepted AI/ML papers. Each diagram is paired with its caption, paper title, and abstract, with citing paragraphs included when available. We combine multi-stage filtering, automated checks, and human audits, with an estimated diagram-selection precision of 93.67%. Retrieval experiments show that citing paragraphs complement captions with information specific to individual diagrams. Under a fixed generation pipeline, replacing PaperBanana's curated reference pool with DiagramBank increases the overall PaperBanana score from 4.62 to 7.88 on PaperBananaBench. Retrieved references also outperform random references on BizGenEval scientific figures. Additional evaluations with different judges and renderers, together with a separate blind human evaluation, demonstrate the utility of DiagramBank for automated diagram generation. Code: https://github.com/csml-rpi/DiagramBank.

cs.IR · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04269

Integrating Machine Learning and Uncertainty Quantification for Real-Time Decision-Making in Digital Twins: A System-Centric Review and Perspective

Yi-Ping Chen, Ying-Kuan Tsai, Seul Lee, Vispi Karkaria, Dex Doksoo Lee, Wei Chen

原文摘要

Digital Twins are evolving from passive virtual representations into closed-loop systems that continuously synchronize with physical assets and drive real-time decisions. Yet predictive modeling, machine learning (ML), uncertainty quantification (UQ), and decision-making have been studied as separate capabilities, leaving limited understanding of their integration and maintenance during online operation. This narrative review, scoped by the functional requirements of real-time decision-making rather than by Digital Twin terminology alone, examines how such decisions are generated and sustained in engineering Digital Twins. We first examine foundational decision-making architectures and representative engineering applications, followed by methods that address their limitations across the decision loop. We identify optimization-based, policy-based, and knowledge-based approaches as the major pathways from virtual representation to action. Across them, ML plays three roles in decision-making: enabling efficient surrogate modeling, transforming heterogeneous physical feedback into decision-relevant information, and accelerating or directly generating decisions. UQ supports uncertainty-aware decisions and model adaptation, while continual validation, drift detection, and adaptation help maintain reliable operation as the physical system evolves. Together, these findings underscore the need for a shift from a model-centric to a system-centric perspective on Digital Twins, where its performance depends on the interactions with sensing, state estimation, computation, decision-making, and adaptation rather than on their standalone performance. Therefore, future research should shift from isolated algorithm development toward system-level co-design, scalable decision-making, and approaches to design, and data-efficient continual adaptation and validation for AI-enabled autonomous systems.

stat.AP · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.00833

Fewer Tokens, More Answers: Verifiable-Reward Tuning of an Ornith-Style VLM for Multi-View Driving QA

T N Rickesh, Pranav Vinodh, Sowmya Kamath S

原文摘要

Reasoning-enabled model derivatives must turn intermediate deliberation into a usable answer within a deployment budget. We study that transition in an Ornith checkpoint adapted to multi-camera driving question answering with verifiable rewards. Paired archived outputs show that, on a selected 400-question development screen with a 256-token cap, GRPO increases answer extraction from 78.25% to 90.25% and accuracy from 55.50% to 63.75%. Of 39 wrong-to-correct transitions, 33 begin with an unextractable base response. The base's 87 unextractable responses all reach the token cap. Checkpoint-verified reconstruction on those same questions gives a much smaller first-pass accuracy gain at 512 tokens: +0.75 percentage points, versus +8.25 at 256. A later answer-cued continuation is analyzed separately; it helps the base more and narrows the adapter difference. On the full 2,511-question development set, this recovery-enabled pipeline reduces mean generated tokens from 108.6 to 47.1, while the paired accuracy interval includes zero gain. We frame these findings as adaptation of answer delivery under finite budgets, with explicit selection and provenance limits. Ornith's self-scaffolding lineage and newer reasoning-enabled derivatives such as Bonsai motivate the problem, but our measured results concern Ornith alone and do not establish a shared architectural mechanism or cross-model transfer.

cs.CV · cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗